Submitted version of the thesis - Airlab, the Artificial Intelligence ...

46 Chapter 4. Control
B.3. The most common used one is the mixed angular and linear movement.
Almost every path was initially implemented with this movement approach.
Others have been used where optimization is needed to cover problems due
to the lack of encoders to control the position of the motors.
4.3 PWM Control
Motors are controlled by applying a PWM signal to the H-Bridges. In order
to control velocity we did not use any encoders, so we don’t have feedback
aboutthemotor rotation, or thedisplacement achieved withtheresultofthe
applied PWM. To have an odometry information without using encoders,
we decided to review and replicate the work done at AIRLab about using
the mice odometry to get feedback from the motors. On the previously
work, a set of mice were reconfigured to obtain the position information,
and calculations were made from these readings to find the odometry. The
working configuration was able get a feedback when the distance between
the mouse sensors and lens was about 1mm, and the distance between lens
and ground was also 1 mm, but this distance is too small for our robot. In
order to increase the distance to 3 cm which is the distance between the
ground and the chassis of the robot, we experimented changing the lens of
the mice.
The mice work like a camera, and the image is captured by the sensor
inside it through a lens placed in front of it. We tested lenses with different
focal lengths and different orientations in order to achieve the 3 cm configuration.
During the experiments, we obtained different performance on
different grounds and different lighting conditions. We tried several illumination
alternatives, such as laser, led, and table lamp. We were able to find
the correct illumination using the led with a wave length readable from the
sensor. And for the lens we used a configuration with a focal length of 10
mm, leaving the distance between the lens and the sensor at 15 mm and
lens to ground at 30 mm. Instead of using encoders that each cost around
50 euro, the possible configuration costs around 5 euro, but the work is not
finished yet and we decided to include this on the robot later.
Since there is no information on the motor speed and displacement of
the robot, the control process mainly depends on vision, and it is not very
precise. There are situations where we cannot determine whether the motor
is moving. When the robot is stuck in a place and the sensors do not detect